Upload
erjon-mamoci
View
225
Download
0
Embed Size (px)
Citation preview
7/29/2019 MreBmrbe
1/8
Structural/functional homology between the bacterial andeukaryotic cytoskeletonsLinda A Amos1, Fusinita van den Ent2 and Jan Lowe3
Structural proteins are now known to be as necessary for
controlling cell division and cell shape in prokaryotes as they are
in eukaryotes. Bacterial ParM and MreB not only have atomic
structures that resemble eukaryotic actin and form similar
filaments, but they are also equivalent in function: the assembly
of ParM drives intracellular motility and MreB maintains the
shape of the cell. FtsZ resembles tubulin in structure and in its
dynamic assembly, and is similarly controlled by accessory
proteins. Bacterial MinD and eukaryotic dynamin appear to have
similar functions in membrane control. In dividing eukaryotic
organelles of bacterial origin, bacterial and eukaryotic proteins
work together.
Addresses
MRC Laboratory of Molecular Biology, Hills Road, CambridgeCB2 2QH, UK1e-mail: [email protected]: [email protected]: [email protected]
Current Opinion in Cell Biology 2004, 16:18
This review comes from a themed issue on
Cell structure and dynamics
Edited by John A Cooper and Margaret A Titus
0955-0674/$ see front matter
2003 Elsevier Ltd. All rights reserved.
DOI 10.1016/j.ceb.2003.11.005
Abbreviations
AMPPNP 50-adenylyl-imidodiphosphate
EM electron microscopyGFP green fluorescent proteinGTP guanosine 50-triphosphateMAPs microtubule associated proteins
MD mitochondrion-dividing
PD plastid-dividingPH pleckstrin homology
Introduction
Almost all eukaryotes depend on polymers of actin andtubulin to organize their cytoplasm, although the same
basic elements are used in a variety of ways [1]. The
ancestors of these two self-assembling dynamic filamen-
tous proteins are believed to have evolved first in prokar-
yotes. The hypothesis that the bacteria cytoskeleton is
related to the eukaryote cytoskeleton was first established
when the bacterial Z-ring, which constricts the cell during
division, was visualized using fluorescence (Figures 1a,b)
and was shown to consistof FtsZ, a protein with a fold that
mirrors tubulin [2] and displays similar dynamic proper-
ties [3,4]. The relationship between eukaryotic and
prokaryotic filaments became pronounced when the pro-
teins MreB, Mbl and ParM [5,6,79] were shown to
possess structures (Figures 2 and 3) and dynamic proper-
ties [7,9] that resemble those of actin.
The ancestors of other eukaryotic cytoskeletal proteins are
likely to be identified in prokaryotes in the near future.
Owing to low sequence conservation, uncovering distant
relationships between structural proteins will depend on
comparisons of their macromolecular and atomic struc-tures, as well as their functional mechanisms; FtsZ amino
acid sequences show less than 20% identity to tubulin,
while MreB and ParM have 15% and 12% identity to actin,
respectively. Furthermore, prokaryotes seem to be more
diverse than eukaryotes because the only cell division
protein to be identified as common to almost all prokaryotic
organisms is FtsZ; however, other common elements
might have been missed because of sequence variation.
Determination of cell shapeActin filaments and microtubules usually cooperate in
controlling the shapes of eukaryotic cells [1]. Cell mem-
branes tend to be supported by an underlying layer of one
or other of these cytoskeletal filaments. Such a role foractin-like filaments seems to have originated in bacteria.
Two proteins, MreB and Mbl (MreB-like), each form
cables that follow a helical path close to the membrane ofBacillus subtilis (Figure 1d) [5,9]. Equivalent proteins
are found in most other non-spherical bacteria. Mutants
lacking MreB are round instead of rod-shaped; however,
some non-spherical bacteria that naturally lack actin-like
proteins control their shape by extending the cell-wall
only at the poles [10].
Investigation of the structure of MreB from Thermotoga
maritima by X-ray crystallography and electron micro-
scopy have amply confirmed that it is an actin homologue
[6]. In the crystal structure (Figure 3), the four subdo-
mains around the central nucleotide binding site in each
monomer show a strong similarity to those of actin. Moreimportantly, MreB was found to crystallize in such a way
that monomers were lined up to form filaments whose
longitudinal contacts confirm predicted contacts in an
actin filament. The 5.1 nm longitudinal spacing of sub-
units in MreB filaments, compared with 5.5 nm for F-actin, was confirmed in electron micrographs of sheets
and filaments that were assembled from purified protein
(Figure 2) MreB appears to form two-stranded filaments
(Figure 2c) similar to F-actin, except that the strands do
not twist around each other. The double-stranded fila-
ments further associate into pairs and larger bundles.
1
COCEBI 114
www.current-opinion.com Current Opinion in Cell Biology 2004, 16:18
7/29/2019 MreBmrbe
2/8
MreB can form curved, as well as straight, bundles in vitro(Figure 2a), suggesting that a cooperative conformational
change can occur.
Chromosome separationIn eukaryotes, pairs of chromosomes are always moved
apart by a spindle that is constructed from tubulin-contain-
ing microtubules and is driven by microtubule-associated
motors. This is the most obvious feature that universally
differentiates eukaryotes from prokaryotes; however,
there are indications that bacteria do have simple cyto-
skeletal machines to assist in DNA segregation.
ParM self-assembly drives plasmid separation
The protein ParM [7] forms highly dynamic actin-like
filaments in the bacterium Escherichia coli and uses a
self-assembly mechanism for force production and trans-
port that is also used by eukaryotic actin. The crystal
structure of the ParM monomer [8] has a clear homol-
ogy to those of actin and MreB (Figure 3). A bonus result
from this study was a crystal structure of ParM without
bound nucleotide. Of the eukaryotic proteins, only the
Figure 1
(a) (b)
(c)
(d)
(e)
Light microscopy of fluorescently-labeled cytoskeletal filaments in
bacteria. (a) Visualization of the Z-ring in E. coli cells expressing GFP
FtsZ; two different views of a cell undergoing constriction. Reprinted
from [17], Copyright (1996) National Academy of Sciences, U.S.A. Bar
1 mm. (b) Brightfield and fluorescence images of an E. coli cell
expressing a low level of GFPFtsZ. Photobleaching experiments
revealed that the Z-ring exchanges subunits rapidly with a cytoplasmic
pool. Reprinted from [4], Copyright (2002) National Academy of
Sciences, U.S.A. Bar 2 mm. (c) Combined phase-contrast andimmunofluorescence microscopy of fixed E. coli cells expressing
wild-type levels of ParM. The fluorescent staining shows intracellular
actin-like filaments. Reprinted from [7], Copyright (2003), with
permission from OUP. (d) Localization of GFPMbl (MreB-like protein)
expressed in Bacillus subtilis shows dynamic helical bands.
Immunofluorescence microscopy of endogenous Mbl in fixed cells gave
similar images. Reprinted from [9], Copyright (2003), with permission
from Elsevier. Bar 4 mm. (e) E. coli cells expressing YFP (yellow
fluorescent protein)-MinD. After deconvolution, these optically sectioned
images of fixed cells reveal that MinD is present in helical filaments.
MinEGFP gave similar images. Reprinted from [47], Copyright (2003)
National Academy of Sciences, U.S.A.
Figure 2
(a)
(b) (c) (d)
Current Opinion in Cell Biology
Electron micrographs of negatively stained filaments. Both MreB [6] and
ParM [8] assemble into protofilaments with a subunit repeat close to
that in F-actin. MreB protofilaments associate into pairs but do not twist
to form helices like actin filaments; the 2-stranded filaments further
associate into straight or curved bundles (a-c); the various MreB
polymers all assemble together in the same conditions. (b) Individual
protofilaments, equivalent to those seen in atomic detail in the crystal
structure of MreB, can also assemble into flat sheets. (c) A pair of
double-stranded filaments; the inset is an enlarged filtered image. (d)
ParM double protofilaments have an even more remarkable structural
similarity to F-actin, as is clear from the enlarged filtered image (right).
Bars 100 nm.
2 Cell structure and dynamics
Current Opinion in Cell Biology 2004, 16:18 www.current-opinion.com
7/29/2019 MreBmrbe
3/8
actin-related protein ARP3 has been seen in the emptyform [11]. In both cases, the empty cleft is 258 wider
than in the ADP-filled form of the protein, indicating a
hinge-like movement between domains. Electron micro-
scopy of assembled filaments showed them to be double-
stranded and helical (Figure 2d), with a subunit arrange-
ment and longitudinal repeat similar to F-actin [8]. The
longitudinal repeat of 4.9 nm is slightly smaller than that
of MreB (5.1 nm) or actin (5.5 nm); as a result of this and
the fact that there are 12.5 monomers per repeat (as
opposed to 12.514 in F-actin), the distance between
crossovers is only 30 nm.
ParM is a product of one of the genes on the R1 plasmid,
which maintains its presence in the bacterial host by
segregating copies of itself towards the two cell poles,
before cell division. Plasmids of this type encode a
centromere-like DNA sequence, called parC in the case
of plasmid R1, and two proteins, called ParM and ParR.Duplicated plasmids are paired at their centromeric re-
gions through direct interactions with ParR. This ParR
parC complex interacts with ParM and stimulates its
ATPase activity. Immunofluorescence microscopy has
revealed ParM-containing filamentous bundles (Figure 1c)
along the longitudinal axis of E. coli[7].
Separation of bacterial chromosomes might be assisted in
a similar way; orthologues of some of the par genes have
been identified on chromosomal DNA (see Update).However, a DNA replication factory at the centre of the
cell [12] is thought to supply at least some of the force that
drives newly duplicated DNA towards opposite cell poles.
Segregation starts with the replication origins, which move
further from the centre as replication continues.
In eukaryotes, several types of protein complexes enable
actin assembly to drive the movement of membranous
organelles or to push forward the leading edge of the cell
membrane [13]. Formins [14] seem to represent the
closest system to the ParRparC complex, in that they
promote the assembly of a simple bundle of actin fila-ments. In such cases, assembly-driven activity might
follow the actoclampin model [15], in which each growing
filament is continuously tethered to a clamp molecule via
a site on the side of the endmost, ATP-bound monomer.
Addition of a new monomer on to the growing end triggersa cycle of ATP-hydrolysis and advancement of the clamp
molecule to the new subunit on the same filament. By
contrast, complexes that incorporate the actin-family pro-
teins ARP2/3 nucleate new forwards-pointing actin fila-
ments as branches from existing filaments [16]. The new
filaments must then make new contacts with the mem-
brane that is being pushed, using new ARP2/3-containingcomplexes.
Although MreB and Mbl helical bundles are also highly
dynamic in vivo [9], it is not known whether their
assembly can drive movement of other objects, in the
manner of ParM or actin. Nor is it known whether any ofthe bacterial actin-like filaments can serve as tracks for
motor protein-driven movement. As yet there is no evi-
dence for linear motor proteins in prokaryotes, apart from
those that move DNA [12].
Cytokinesis/septationConstriction and abscission in different cell types
The division of most cells involves constriction of the
membrane until there is a narrow connection between
the daughter cells (Figure 4) and this is followed by a
separate abscission step. One common theme in both
Figure 3
Current Opinion in Cell Biology
Ribbon structures of actin-family monomers: Actin.ADP, from pdb 1J6Z;
eukaryotic actin-related protein ARP3, from pdb 1K8K [11]; bacterial
ParM.ADP, from pdb 1MWM; ParM empty, from pdb 1MWK [8];
bacterial MreB.AMPPNP, from pdb 1JCG [6]. Each monomer has two
similar domains with a nucleotide binding site in the cleft between them.
The four subdomains (1A, 1B, 2A, 2B) in each protein are coloured blue,
yellow red and green. The crystal structure of ParM with and without
ADP [8] indicates a domain movement of 258, close to the
conformational change predicted for actin from a comparison of
actin.ADP with ARP3 in the empty state and from EM studies of F-actin
under different conditions [67]. The protofilament in the MreB crystal
structure provided the first atomic-resolution view of inter-subunit
contacts in an actin-family filament [6].
Structural/functional homology between the bacterial and eukaryotic cytoskeletons Amos, van den Ent and Lowe 3
www.current-opinion.com Current Opinion in Cell Biology 2004, 16:18
7/29/2019 MreBmrbe
4/8
steps seems to be the use of filamentous GTP-binding
proteins (FtsZ, septins, dynamin), although they are notclosely related.
In bacteria, the Z-ring that is assembled from FtsZ
initially defines the site of the division plane [1719]
and is then actively involved in the constriction mechan-
ism. Other proteins are recruited to the ring, depending
on species. In E. coli and B. subtilis, these include FtsA,
another member of the actin family, and ZipA. In the
early stage of constriction, either ZipA or FtsA can sta-
bilize the Z-ring but FtsA is required for the final abscis-
sion [20,21]. Several other proteins are required to
coordinate changes in the cell membrane with formation
of a septum in the cell wall [18].
In fungi and animal cells, but apparently not in plants, the
division plane is first defined by a ring assembled from
septin filaments [22,23]. Myosin is then recruited [24] andactin bundles are assembled, organized by formins [14].
Acto-myosin sliding-filament contraction constricts the
cell diameter until it is filled by the central-spindle
bundle of microtubules (Figure 4).
In higher plants, the division plane is first defined by a
dynamic ring of microtubules [25] but these do not cause
constriction. Subsequently, other microtubules assemble
at right angles to the division plane, forming a phragmo-
plast [26], which is similar to an extended central spindle.The final stage of cytokinesis is, perhaps, most clear in
higher plants; Golgi-derived vesicles are transported
along the microtubules of the phragmoplast to the plane
of division, where they fuse with one another to form the
incipient cell wall. Members of the dynamin family are
essential components here [27] and also in animal cells
[28]; in this situation, the role of dynamin could be either
to package new membrane into tubes that can be deliv-
ered to the new cell wall and/or to remodel the newly
fused membrane.
Dynamic assembly of FtsZ and its control
FtsZ and tubulin both assemble into protofilaments with
a subunit spacing of 4.04.2 nm [29]. GTP, bound to a
pocket at the top of one subunit, comes into contact with
the bottom of the next subunit where a loop, known as
T7, on the latter surface is responsible for triggering GTPhydrolysis [30]. In the case of tubulin heterodimers, the
T7 loop ofa-tubulin is thought to hydrolyse GTP that is
bound to the b-tubulin in another dimer, but the T7 loop
ofb-tubulin simply traps the GTP that is bound to its own
a-tubulin partner. FtsZ might form homodimers as inter-
mediate complexes during assembly [31] but GTP is,
presumably, hydrolysed in every monomer. Both tubulinand FtsZ polymers become unstable after GTP hydro-
lysis and exhibit dynamic instability. In vivo, the Z-ring is
even more dynamic than microtubules [4]. The rapid
turnover of GTP means that it is difficult to study the
assembly of purified FtsZ in vitro, although some progress
has been made recently [3234]. However, the preciseform of the in vivo polymer remains unknown. It is
unlikely to closely resemble eukaryotic microtubules
because lateral contacts between tubulin protofilaments
are made by polypeptide loops that are missing from FtsZ
[2]. It might consist of just a pair of protofilaments [29,33].
Just as microtubule stability is controlled by microtubule
associated proteins (MAPs), assembly of FtsZ is pro-
moted in vivo by proteins such as FtsA, ZipA [1921]
and ZapA [35]. ZipA, which has a long unfolded P/Q rich
domain [36], perhaps best resembles typical MAPs, which
Figure 4
SeptinsMyosin IIFormins
Actin
AnillinSpectrin
DynaminSyntaxin
MTsMTs
FtsZFtsAZipAZapAMreB MreB
ParM ParM
MinCDMinCD
FtsB, FtsIFtsK, FtsLFtsN, FtsQ
FtsW
(a)
(b)
Current Opinion in Cell Biology
Some of the cytoskeleton proteins involved in cytokinesis. (a)
Representation of a section through an animal cell when the actin-
containing cleavage ring has contracted around the central bundle of
microtubules (MTs) of the extended mitotic spindle, after the
chromosomes have separated. Several proteins aside from actin are
required to form the cleavage ring and produce constriction. Dynamin
and syntaxin are required for membrane remodeling during the final
stage of separation. (b) Representation of a section through a dividing
bacterium. Proteins FtsA, ZipA and ZapA join the tubulin homologue
FtsZ in the Z-ring, which constricts the cell membrane during division.
Mutations in other Fts (filamenting temperature-sensitive) proteins
prevent invagination of the outer cell wall. The actin homologue, ParM,
separates copies of the plasmid on which it is encoded. Actin-like MreB
is needed to give the cells the correct shape and also to segregate the
chromosomes correctly. The Min proteins are needed to prevent a Z-ring
from forming too close to a pole.
4 Cell structure and dynamics
Current Opinion in Cell Biology 2004, 16:18 www.current-opinion.com
7/29/2019 MreBmrbe
5/8
are unstructured in the absence of tubulin. MinC appears
to inhibit the assembly of unwanted Z-rings (discussed in
the next section) by binding to FtsZ polymer (and dis-
placing FtsA [31]). The structure of the MinC dimer is
compatible with a model in which it binds to the sides of
two adjacent subunits in a protofilament [37], possiblyinhibiting a lateral association of protofilaments that
might be required to form the Z-ring [38]. Alternatively,
MinC might cause disassembly by inducing curvature in
FtsZ protofilaments. There is a precedent for this second
mechanism, in the way that stathmin interacts with two
tubulin dimers [39]. However, there is no evidence that
MinC sequesters FtsZ in the way that stathmin seques-
ters tubulin. When E. coli or B. subtilis are under stress,
FtsZ is sequestered by a protein called SulA, which
prevents GTP hydrolysis by binding to the T7 loop[40], or YneA [41], which may perform a similar role.
The Min systems in bacteria
A protein complex, called MinCD, is thought to be
responsible for inhibiting FtsZ from assembling prema-
turely on to sites that will ultimately be at the centres of
the two future daughter cells. MinD binds cooperatively
to membranes in the presence of ATP and recruits MinC,
which is required for the regulation of Z-ring assembly
[31,37]. In B. subtilis, a protein called DivIVA recruits
MinCD to each polar region [42], leaving only the site at
the centre of the parent cell free to assemble a Z-ring but,in E. coli, MinCD is not localized in this way. Instead, a
soluble protein, MinE, prevents MinCD from inhibiting
Z-ring assembly at the central site [43,44,45,46]. The
activity of MinE sets up a surprising oscillation of all
the Min proteins. Originating from the cell centre, a ringof MinE progresses towards one pole, disassembling
MinCD as it moves. The depolymerized MinC and MinD
are then free to assemble near to the other pole. On
reaching the first pole, MinE is released to work on
MinCD polymers that are formed in the second half of
the cell. Each movement, from one half of the cell to the
other, takes 3050 s. Meanwhile, the whole cell graduallyincreases in length. Recent images with improved resolu-
tion [47] have shown that MinCD does not coat the
whole cell membrane near one pole but forms a long
winding filament (Figure 1e); its relationship to the MreB
filaments, if any, is still unclear.
MinD resembles dynamin
The similarities between dynamin and MinD might be
purely coincidental, but it is interesting to compare them
because they could share a common mechanism. The N-
terminal domain of dynamin is a regular GTPase, with a
strong similarity to Ras [48]. GTP regulates dimerization,
as well as interactions with membranes and with a variety
of actin-binding proteins [27,28]. The other domains
have important roles in these interactions [49]. A 3 D
reconstruction of a dynamin polymer surrounding a mem-
brane tubule has been calculated from electron micro-
scope images [50]. It shows a helical arrangement of
dimeric GTPase domains, which stand out from the
membrane on legs, consisting of the smaller, regulatory
domains. Contact with the membrane is via PH (pleck-
strin homology) domains at the ends of the legs.
The structure of MinD has been solved in the empty
state as well as when bound to the nonhydrolysable
ATP analogue AMPPNP (50-adenylyl-imidodiphosphate)
[5153]. It belongs to a group of ATPases that are related
to the P-loop GTPases [54]. The role of ATP is to
modulate the interaction of the protein with itself and
its partners, MinC, MinE and membranes [43,44,
45,46]. In vitro, MinD is known to dimerize with ATP
before assembling directly on to the membrane to form a
fairly tight helical polymer [43,44,45]. This polymer issuperficially reminiscent of the outer wall of the dynamin
helix, but presumably has no legs distancing it from the
membrane inside. Interaction with the membrane is via a
short C-terminal segment [55,56], which is much smaller
than the multidomain region of dynamin that follows the
GTPase domain.
One possibility is that the MinCD polymer that is
assembled in vivo (Figure 1) contains a tubule of fresh
membrane that becomes incorporated into the cell mem-
brane when released by MinE; such a mechanism would
help ensure an even insertion of new lipid along thelength of the cell. This would be similar to the proposed
role for dynamin in organizing membrane insertion during
the abscission of eukaryotic cells [27,28].
Division of chloroplasts and mitochondriaFtsZ, MinD and MinE are found in eukaryotes in
chloroplasts (derived from cyanobacteria) [57] and in
some primitive mitochondria (derived from a-proteobac-
teria) and they serve important functions in organelle
division [58,59]. FtsZ is presumably required for mark-
ing the division site and/or constricting the inner mem-
brane. Eukaryotic proteins, such as actin, organelle-specific dynamins [6061,62] and septins [63], are also
necessary for organelle division. Thus, the division
mechanism is a hybrid of prokaryotic constriction and
eukaryotic abscission. Most mitochondria lack FtsZ and,
apparently, use only eukaryotic division mechanisms.
Electron microscopy of chloroplast structure [64] shows
electron-dense plastid-dividing (PD) rings, on the cyto-
plasmic face of the outer membrane and on the stromal
face of the inner membrane. Mitochondria have a similar
structure, known as the mitochondrion-dividing (MD)
ring [65]. The chloroplast FtsZ ring localizes on the
stromal side of the inner PD ring. Both the Z-ring and
then the PD rings appear before division starts but the Z-
ring disappears before the final abscission takes place. An
attempt to isolate the PD rings in detergent [64] yielded
insoluble 5 nm filaments of unknown composition (but
Structural/functional homology between the bacterial and eukaryotic cytoskeletons Amos, van den Ent and Lowe 5
www.current-opinion.com Current Opinion in Cell Biology 2004, 16:18
7/29/2019 MreBmrbe
6/8
the absence of septins from plants might be worth rein-
vestigating [63]). A chloroplast-specific dynamin associ-
ates with the chloroplast outer membrane at a late stage of
division [60,61] and it might be of significance that
this is when the outer membrane needs to increase in
area, even more than the inner membrane.
The activity of the Z-ring has been more clearly tracked
in chloroplasts than in any bacterium, although we cannot
be certain that the constriction mechanism, in conjunc-
tion with that of eukaryotic mechanisms, is truly repre-
sentative of that in bacteria.
ConclusionsCytoskeletal proteins that are related in both structure
and function are currently being identified in bacteria andeukaryotic cells. The actin-like and tubulin-like proteins
in bacteria are evolutionarily related to their eukaryotic
counterparts, having similar functions and 3D structures.
Bacteria also have accessory proteins that control FtsZ
assembly in the way that MAPs control tubulin assembly,
although there is no evidence that any are evolutionary
precursors of MAPs. As yet, no associated proteins for
bacterial actin have been found; the search for them will
be an important objective in the near future. Researchers
will also continue to look for prokaryotic motor proteins
that might use FtsZ or bacterial actin filaments as tracks.
The role of dynamins in the division of chloroplasts andmitochondria, and their role in eukaryotic cytokinesis, is
another exciting area for further research. Finally, the full
role of MinD and its accessory proteins in bacteria and
chloroplasts demands declarative investigation.
UpdateRecent immunofluorescence studies of helical MreB fila-
ments in E. coli [66] have demonstrated that chromo-
somes fail to segregate properly in the round cells formed
when MreB is missing or mutated. These results suggest
that there is a bacterial chromosome segregation mechan-
ism that is homologous to active plasmid separation [7].
References and recommended readingPapers of particular interest, published within the annual period ofreview, have been highlighted as:
of special interest
of outstanding interest
1. Nelson WJ: Adaptation of core mechanisms to generate cellpolarity. Nature 2003, 422:766-774.
2. Nogales E, Downing KH, Amos LA,LoweJ: Tubulinand FtsZ forma distinct family of GTPases. Nat Struct Biol 1998, 5:451-458.
3. Mukherjee A, Lutkenhaus J: Analysis of FtsZ assembly by lightscattering and determination of the role of divalent metalcations. J Bacteriol 1999, 181:823-832.
4.
Stricker J, Maddox P, Salmon ED, Erickson HP: Rapid assemblydynamics of the Escherichia coli FtsZ-ring demonstrated byfluorescence recovery after photobleaching. Proc Natl Acad SciUSA 2002, 99:3171-3175.
The E. coli Z-ring was visualized with green fluorescent protein-labeledFtsZ.After photobleaching,the fluorescencerecoveredwith a half-time of
30 s. The rapid turnover correlates well with the high GTPase activity ofFtsZ in vitro. A mutant FtsZ, whose activity is much reduced, showed acorresponding reduction in turnover. Thus, the assembly dynamics ofFtsZ, like those of tubulin, are regulated by GTP hydrolysis. Interestingly,the mutant cells divided with a normal cell-cycle time.
5. Jones LJF, Carballido-Lopez R, Errington J: Control of cell shapein bacteria: helical, actin-like filaments in Bacillus subtilis.
Cell 2001, 104:913-922.6. van den Ent F, Amos LA, Lowe J: Prokaryotic origin of the actin
cytoskeleton. Nature 2001, 413:39-44.
7.
Mller-Jensen J, Jensen RB, Lowe J, Gerdes K: ProkaryoticDNA segregation by an actin-like filament. EMBO J 2002,21:3119-3127.
The first evidence that prokaryotic DNA separation employs a simpleanalogue of the eukaryotic mitotic spindle apparatus. R1, an E. coliplasmid, codes for ParM, an actin-like ATPase. Here, polymerization ofParM is shown to generate the necessary force for partitioning copies ofthe plasmid to opposite cell poles. Filament formation was dependent onthe other components encoded by the plasmid (i.e. protein ParR andthe centromere-like parC region of the plasmid DNA, to which ParRbinds); the ParRparC complex functions as a nucleation point for ParMpolymerization.
8.
van den Ent F, Mller-Jensen J, Amos LA, Gerdes K, Lowe J:F-actin-like filaments formed by plasmid segregation protein
ParM. EMBO J 2002, 21:6935-6943.Crystal structures of protein ParM show a domain movement of 258 onbinding of ADP. The protein is also shown to polymerize into doublehelical filaments with a longitudinalsubunit repeatthat is similar to F-actinand to the MreB filaments that maintain the cell shape of non-sphericalbacteria.
9.
Carballido-Lopez R, Errington J: The bacterial cytoskeleton: invivo dynamics of the actin-like protein Mbl of Bacillus subtilis.Dev Cell 2003, 4:19-28.
Immunofluorescence studies using antibodies against actin-like MreBand Mbl proteins first showed filamentous helical structures close to thecell membrane. MreB and Mbl were shown to be required for controllinglateral and longitudinal cell dimensions, respectively.
10. Daniel RA, Errington J: Control of cell morphogenesis inbacteria: two distinct ways to make a rod-shaped cell.Cell 2003, 113:767-776.
11. Robinson RC, Turbedsky K, Kaiser DA, Marchand JB, Higgs HN,
Choe S, Pollard TD: Crystal structure of Arp2/3 complex.Science 2001, 294:1679-1684.
12. Sherratt DJ: Bacterial chromosome dynamics. Science 2003,301:780-785.
13. Pantaloni D, Clainche CL, Carlier M-F: Mechanism of actin-basedmotility. Science 2001, 292:1502-1506.
14. Waller BJ, Alberts A: S: The formins: active scaffolds thatremodel the cytoskeleton. Trends Cell Biol 2003, 13:435-446.
15. Dickinson RB, Purich DL: Clamped-filament elongation modelfor actin-based motors. Biophys J 2002, 82:605-617.
16. Pollard TD, Borisy GG: Cellular motility driven by assembly anddisassembly of actin filaments. Cell 2003, 112:453-465.
17. Ma X, David W, Ehrhardt DW, Margolin W: Colocalization of celldivision proteins FtsZ and FtsA to cytoskeletal structures inliving Escherichia colicells by using green fluorescent protein.
Proc Natl Acad Sci USA 1996, 93:12998-13003.
18. Buddelmeijer N, Beckwith J: Assembly of cell divisionproteins at the E-coli cell center. Curr Opin Microbiol 2002,5:553-557.
19. Rueda S, Vicente M, Mingorance J: Concentration and assemblyof the division ring proteins FtsZ, FtsA, and ZipA during the
Escherichia coli cell cycle. J Bacteriol 2003, 185:3344-3351.
20. Pichoff S, Lutkenhaus J: Unique and overlapping roles for ZipAand FtsA in septal ring assembly in Escherichia coli.EMBO J 2002, 21:685-693.
21. Geissler B, ElrahebD, Margolin W:A gain-of-function mutation inftsA bypasses the requirement for the essential cell divisiongene zipA in Escherichia coli. Proc Natl Acad Sci USA 2003,100:4197-4202.
6 Cell structure and dynamics
Current Opinion in Cell Biology 2004, 16:18 www.current-opinion.com
7/29/2019 MreBmrbe
7/8
22. Longtine MS, Bi E: Regulation of septin organization andfunction in yeast. Trends Cell Biol 2003, 13:403-409.
23. Macara IG, Baldarelli R, Field CM, Glotzer M, Hayashi Y, Hsu SC,Kennedy MB, Kinoshita M, Longtine M, Low C et al.: Mammalianseptins nomenclature. Mol Biol Cell 2002, 13:4111-4113.
24. Shu S, Liu X, Korn ED: Dictyostelium and Acanthamoebamyosin II assembly domains go to the cleavage furrow of
Dictyostelium myosin II-null cells. Proc Natl Acad Sci USA 2003,100:6499-6504.
25. Dhonukshe P, Gadella TW: Alteration of microtubule dynamicinstability during preprophase band formation revealed byyellow fluorescent protein-CLIP170 microtubule plus-endlabeling. Plant Cell 2003, 15:597-611.
26. Smith LG: Plant cell division: building walls in the right places .Nat Rev Mol Cell Biol 2001, 2:33-39.
27.
Kang B-H, Busse JS, Bednarek SY: Members of the Arabidopsisdynamin-like gene family, ADL1, are essential for plantcytokinesis and polarized cell growth. Plant Cell 2003,15:899-913.
Members of the Arabidopsis dynamin-related protein family are shown tobe essential for polarized cell expansion and cytokinesis; mutants showdefects in cell plate assembly andcell wall formation. Thedistributionof afunctional green fluorescent fusion protein is dynamic and localized
asymmetrically on the plasma membrane.28.
Thompson HM, Skop AR, Euteneuer U, Meyer BJ, McNiven MA:The large GTPase dynamin associates with the spindlemidzone and is required for cytokinesis. Curr Biol 2002,12:2111-2117.
In both mammalian and Caenorhabditis elegans cells, dynamin localizesto the spindle midzone and is enriched in spindle midbody extracts. In C.elegans, dynamin also localizes to newly formed cleavage furrow mem-branes. Embryos from a temperature-sensitive mutant and dynaminRNAi-treated embryos were defective in the late stages of cytokinesis,suggesting that dynamin plays an essential role in the final separation ofdividing cells.
29. Lowe J, Amos LA: Tubulin-like protofilaments in Ca2R-inducedFtsZ sheets. EMBO J 1999, 18:2364-2371.
30.
Scheffers DJ, de Wit JG, den Blaauwen T, Driessen AJ: GTPhydrolysis of cell division protein FtsZ: evidence that theactive site is formed by the association of monomers.
Biochemistry 2002, 41:521-529.In mixtures of wild-type FtsZ and proteins with a mutation in the T7 loop(on the opposite side of the molecule from the GTP binding site), mostmutants were found to inhibit wild-typeGTP hydrolysis.This confirmsthata T7 loop is required to complete the active site.
31. Justice SS, Garcia-Lara J, Rothfield LI: Cell division inhibitorsSulA and MinC/MinDblock septum formation at different stepsin the assembly of the Escherichia coli division machinery.Mol Microbiol 2000, 37:410-423.
32. Gonzalez JM, Jimenez M, Velez M, Mingorance J, Andreu JM,Vicente M, Rivas G: Essential cell division protein FtsZassembles into one monomer-thick ribbons under conditionsresembling the crowded intracellular environment. J Biol Chem2003, 278:37664-37671.
33. Oliva MA, Huecas S, Palacios JM, Martin-Benito J, Valpuesta JM,Andreu JM: Assembly of archaeal cell division protein FtsZ anda GTPase-inactive mutant into double-stranded filaments.
J Biol Chem 2003, 278:33562-33570.
34. Caplan MR, Erickson HP: Apparent cooperative assembly of thebacterial cell division protein FtsZ demonstrated by isothermaltitration calorimetry. J Biol Chem 2003, 278:13784-13788.
35. Gueiros-Filho FJ, Losick R: A widely conserved bacterial celldivision protein that promotes assembly of the tubulin-likeprotein FtsZ. Genes Dev 2002, 16:2544-2556.
36. OhashiT, Hale CA,de Boer PAJ, Erickson HP: Structural evidencethat the P/Q domain of ZipA is an unstructured, flexible tetherbetween the membrane and the C- terminal FtsZ-bindingdomain. J Bacteriol 2002, 184:4313-4315.
37. Cordell SC, Anderson RE, Lowe J: Crystal structure of thebacterial cell division inhibitor MinC. EMBO J 2001,20:2454-2461.
38. JohnsonJE, Lackner LL,de Boer PAJ: Targeting of (D)MinC/MinDand (D)MinC/DicB complexes to septal rings in Escherichiacoli suggests a multistep mechanism for MinC-mediateddestruction of nascent FtsZ rings. J Bacteriol 2002,184:2951-2962.
39. Gigant B, Curmi PA, Martin-Barbey C, Charbaut E, Lachkar S,Lebeau L, Siavoshian S, Sobel A, Knossow M: The 4A- x-ray
structure of a tubulin:stathmin-like domain complex. Cell2000,102:809-816.
40.
Cordell SC, Robinson EJH, Lowe J: Crystal structure of the SOScell division inhibitor SulA and in complex with FtsZ.Proc Natl Acad Sci USA 2003, 100:7889-7894.
SulA is shown to inhibit FtsZ assembly by binding to the T7 loop surface,which forms part of an active site when FtsZ is assembled into proto-filaments. The conformation of FtsZ in the complex with SulA is the sameas in crystals of FtsZ alone.
41. Kawai Y, Moriya S, Ogasawara N: Identification of a protein,YneA, responsible for cell division suppression during the SOSresponse in Bacillus subtilis. Mol Microbiol 2003, 47:1113-1122.
42. Harry EJ, Lewis PJ: Early targeting of Min proteins to the cellpoles in germinated spores of Bacillus subtilis: evidence fordivision apparatus-independent recruitment of Min proteins tothe division site. Mol Microbiol 2003, 47:37-48.
43.
Suefuji K, Valluzzi R, RayChaudhuri D: Dynamic assembly, of
MinD into filament bundles modulated by ATP, phospholipids,and MinE. Proc Natl Acad Sci USA 2002, 99:16776-16781.
This paper, together with [32,33,34], confirms the importance of themembrane as a partner in interactionsbetween theMin proteins,Min C, Dand E. EM images of MinD filaments assembled in the presence of ATPare shown here.
44.
Hu Z, Gogol EP, Lutkenhaus J: Dynamic assembly of MinD onphospholipid vesicles regulated by ATP and MinE. Proc NatlAcad Sci USA 2002, 99:6761-6766.
Includes an impressive EM image of MinD bound to phospholipidvesiclesandassembled into a helical array that deformed thevesiclesinto tubes inthe presence of ATP. Addition of MinE stimulated MinDs ATPase andcaused its release from the vesicles.
45. Lackner LL, Raskin DM, de Boer PAJ: ATP-dependentinteractions between Escherichia coli Min proteins andthe phospholipid membrane in vitro. J Bacteriol 2003,185:735-749.
46.
Hu Z, Saez C, Lutkenhaus J: Recruitment of MinC, an inhibitor ofZ-ring formation, to the membrane in Escherichia coli: role ofMinD and MinE. J Bacteriol 2003, 185:196-203.
A more detailed study. In the absence of lipid, ATP causes MinD todimerize and form a rather unstable complex with MinC. However, MinDbinds well to lipid vesicles in the presence of ATP and then recruits MinC.If MinD release via ATP hydrolysis is prevented, MinE can displace MinCthat is bound to the MinD-lipid complex but MinC does not displacebound MinE.
47.
Shih Y-L, Le T, Rothfield L: Division site selection in Escherichiacoli involves dynamic redistribution of Min proteins withincoiled structures that extend between the two cell poles .Proc Natl Acad Sci USA 2003, 100:7865-7870.
Imagesoffluorescently-labeled proteins showthat the threeMin proteins,all known to oscillate between the two poles of E. coli, are dynamicallyassociated with helices that wind around the cytoplasmic side of the cellmembrane. The authors also show that the E. coli MreB protein formsextended helices similar to those previously reported for B. subtilis [7].
TheMreB andMinCDE coiledarrays do notappear identicalbut this pointneeds to be confirmed by co-labeling the proteins in the same cells.
48. Niemann HH, Knetsch MLW, Scherer A, Manstein DJ, Kull FJ:Crystal structure of a dynamin GTPase domain in bothnucleotide-free and GDP-bound forms. EMBO J 2001,20:5813-5821.
49. Hinshaw JE: Dynamin and its role in membrane fission.Annu Rev Cell Dev Biol 2000, 16:483-519.
50. Zhang P, Hinshaw JE: Three-dimensional reconstruction ofdynamin in the constricted state. Nat Cell Biol 2001, 3:922-927.
51. Hayashi I, Oyama T, Morikawa K: Structural and functionalstudies of MinD ATPase: implications for the molecularrecognition of the bacterial cell division apparatus. EMBO J2001, 20:1819-1828.
Structural/functional homology between the bacterial and eukaryotic cytoskeletons Amos, van den Ent and Lowe 7
www.current-opinion.com Current Opinion in Cell Biology 2004, 16:18
7/29/2019 MreBmrbe
8/8
52. Cordell SC, Lowe J: Crystal structure of the bacterial celldivision regulator MinD. FEBS Lett 2001, 492:160-165.
53. Sakai N, Itou H, Watanabe N, Yao M, Tanaka I: The MinD proteinfrom the hyperthermophilic archaeon Pyrococcus horikoshii:crystallization and preliminary X-ray analysis.Acta Crystallogr DBiol Crystallogr 2001, 57:896-897.
54. Leipe DD, Wolf YI, Koonin EV, Aravind L: Classification andevolution of P-loop GTPases and related ATPases.J Mol Biol 2002, 317:41-72.
55. Szeto TH, Rowland SL, Rothfield LI, King GF: Membranelocalization of MinD is mediated by a C-terminal motif that isconserved across eubacteria, archaea, and chloroplasts.Proc Natl Acad Sci USA 2002, 99:15693-15698.
56. Hu ZL, Lutkenhaus J: A conserved sequence at the C-terminusof MinD is required for binding to the membrane andtargeting MinC to the septum. Mol Microbiol 2003,47:345-355.
57. Martin W, Rujan T, Richly E, Hansen A, Cornelsen S, Lins T,Leister D, Stoebe B, Hasegawa M, Penny D: Evolutionary analysisof Arabidopsis, cyanobacterial, and chloroplast genomesreveals plastid phylogeny and thousands of cyanobacterialgenes in the nucleus. Proc Natl Acad Sci USA 2002,99:12246-12251.
58. Vitha S, McAndrew RS, Osteryoung KW: FtsZ ring formationat the chloroplast division site in plants. J Cell Biol 2001,153:111-119.
59.
Reddy MSS, Dinkins R, Collins GB: Overexpression of theArabidopsis thaliana MinE1 bacterial division inhibitorhomologue gene alters chloroplast size and morphology intransgenic Arabidopsis and tobacco plants. Planta 2002,215:167-176.
Confocal and electron-microscopy of plants overexpressing MinE1revealed chloroplasts with abnormal shapes and sizes. This indicatesthat this protein probably has a similar role to bacterial MinE.
60.
Gao H, Kadirjan-Kalbach D, Froehlich JE, Osteryoung KW: ARC5,a cytosolic dynamin-like protein from plants, is part of thechloroplast division machinery. Proc Natl Acad Sci U S A 2003,100:4328-4333.
Unlike most mitochondria, chloroplasts have retained FtsZ for division.This paper (and [61,62]) show that a requirement for a dynamin-related
protein, with no obvious counterparts in prokaryotes, is common to bothtypes of endosymbiotic organelle. ARC5 is related to a group of dynamin-like proteins that are unique to plants. A GFPARC5 fusion proteinlocalizes to a ring at the chloroplast division site and ARC5 mutationscause enlarged, dumbbell-shaped chloroplasts. Import and protease
protection assays indicate that the ARC5 ring is positioned on the outersurface of the chloroplast.
61.
Miyagishima SY, Nishida K, Mori T, Matsuzaki M, Higashiyama T,Kuroiwa H, Kuroiwa T: A plant-specific dynamin-related proteinforms a ring at the chloroplast division site . Plant Cell 2003,15:655-665.
Analysis of synchronized chloroplasts showed the time at which a
dynamin-like protein is recruited to chloroplasts, forming a ring on thecytosolic side of the division site in the late stage of division. The ringconstricts until division is complete and then disappears.
62.
Nishida K, Takahara M, Miyagishima S, Kuroiwa H, Matsuzaki M,Kuroiwa T: Dynamic recruitment of dynamin for finalmitochondrial severance in a primitive red alga. Proc Natl AcadSci USA 2003, 100:2146-2151.
Previously, FtsZ was shown to localize to mitochondria in primitive algaeon the inside surface of the division site. Here, it is shown that FtsZ,dynamin andthe MD ring,an electron-dense structureon thecytoplasmicside of the mitochondrial membrane, act during different phases ofdivision. Before final separation, dynamin patches are recruited exter-nally, to the midpoint of the constricted MD ring.
63.
Takahashi S, Inatome R, Yamamura H, Yanagi S: Isolation andexpression of a novel mitochondrial septin that interacts withCRMP/CRAM in the developing neurones. Genes Cells 2003,8:81-93.
This presence of a septin in mitochondria might mean that septins are
candidates for components of the MD ring [60
] (and also chloroplast PDrings [64]).
64.
Kuroiwa H, Mori T, Takahara M, Miyagishima S, Kuroiwa T:Chloroplast division machinery as revealed byimmunofluorescence and electron microscopy. Planta 2002,215:185-190.
The behaviour of FtsZ and PD rings, both components of the chloroplastdivision machinery,was followed during the course of division.FtsZ formsa ring onthe stromal division site beforeinvagination starts. Then theinner(stromal) and outer (cytosolic) PD rings appear. The FtsZ ring remains atthe leading edge of the constriction but does not change width althoughthe volume of the outer PD ring gradually increases. The FtsZ ringdisappears before the final stage of chloroplast constriction.
65. Kuroiwa T, Kuroiwa H, Sakai A, Takahashi H, Toda K, Itoh R:The division apparatus of plastids and mitochondria. Int RevCytol 1998, 181:1-41.
66. Kruse T, Mller-Jensen J, Lbner-Olesen A, Gerdes K:
Dysfunctional MreB inhibits chromosome segregation inEscherichia coli. EMBO J 2003, 22:5283-5292.
67. Egelman EH:Actins prokaryotic homologs. Curr Opin Struct Biol2003, 13:244-248.
8 Cell structure and dynamics
Current Opinion in Cell Biology 2004, 16:18 www.current-opinion.com